Re: please expand re bidi algorithm

From: Michael \(michka\) Kaplan (
Date: Mon Oct 02 2000 - 22:55:29 EDT

No, its not that at all. It is just that many products have a long history
of connection with the people who use the product who also happen to have a
bidirectional language as their native one. Many other products have a
development team with that expertise.

One example can be found in Mozilla (NN 6.x) -- tested with milestone 16 on
Windows 2000 and NT4. The slash character "/" is in the Unicode Bidi
algorithm classified as a "European Separator" which means that text would
be expected to be reversed (i.e. 1/2 would be expected to be 2/1). However,
the the Farsi language (which uses the Arabic script) the slash is also a
decimal separator for currency, thus 123.45 would appear as 123/45. However,
this will appear as 45/123 with any program that completely implements the
Bidi algorithm. Therefore, it is unsurprising that programs like Mozilla
would choose to instead classify the slash as a Common Separator, which
allows the desired behavior for Farsi. Several other products (including
Microsoft's Uniscribe technology) use the same sort of variation...

But there is a solid reason for there to be such a change, so who is to say
whether the bug is in Unicode's Bidi algorithm or in the actual software
products. The only thing we can be sure of is that the bug is NOT in the
expectations of the users in Iran. :-)

My point in all of this is to say that I guess you COULD claim that this
means that programs that have this bug are "non-compliant" but I tend to
view this as similar to the Smalltalk developers who refuse to accept C++ as
being object oriented. In other words, it would be a case of being too
technical and paying too much attention to the trees, not enough to the

In any case, it means that your original question about Unicode compliance
is pretty much without real meaning. Why? Well, whose job would it be to
check the compliance? And who would pay them to do this? This is really just
not possible. In the end, each company will make its claims and hopefully
none of them will ever be dishonest about it and each will fully investigate
the consequences of such a variation (like I assume this decision about
Farsi was handled). The end result is indeed what a native Farsi speaker
expects. I would find it unfortunate if this "bug" were "fixed" since it
would be done at the expense of Farsi users.

Just my two cents; I have no official idea what the official views of any of
these products are. :-)


a new book on internationalization in VB at

----- Original Message -----
From: "Elaine Keown" <>
To: <>
Cc: <>
Sent: Monday, October 02, 2000 7:27 PM
Subject: please expand re bidi algorithm

> Dear Michka and List:
> Please elaborate on what you said mean there are multiple
technical approaches to bidi, something like that? Do bidi algorithms
have a long history in the world of algorithms?
> > my example about the Unicode Bidi algorithm is hardly contrived, since
> > products are technically non-compliant, but in ways that I tend to agree
> > with the differences and not with the algorithm.
> Elaine
> _______________________________________________________________________
> Free Unlimited Internet Access! Try it now!
> _______________________________________________________________________

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:14 EDT